Criteria for Restoration of Longitudinal Barriers, Phase II (2021)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

340 CHAPTER 13 – ANCHOR STRENGTH QUANTIFIED IN TERMS OF ANCHOR DAMAGE In Chapter 9 the performance of the G4(2W) guardrail system was quantified in terms of anchor strength. In this chapter, anchor strength will be defined in terms of various anchor damage modes and levels of damage. Combining the results of Chapter 9 with the results presented in this chapter will provide a means of correlating “measureable” damage modes directly to a corresponding level of degradation in guardrail performance. Recall from Chapter 3 that essentially all end-terminals, with the exception of the buried-in-backslope end-terminal, share the same basic features and components which function to ensure that the guardrail maintains proper tension during impacts. The basic features of these systems were presented earlier and will not be repeated here. The focus of this chapter is on the development of procedures for assessing damage modes that directly affect the anchor strength of end-terminals. NCHRP Report 656 provides assessment criteria for generic end-terminals based on a “check list for repairing energy absorbing end terminals” developed by the Ohio Department of Transportation.[Gabler10] The assessment criteria in Report 656 covered damaged end-posts, missing or slack anchor cable, improper stub height, missing or failed lag bolts on impact head, and missing or misaligned bearing plate. A summary of generic end terminal repair guidance from Report 656 is shown in Table 76. These criteria were “based solely on engineering judgment; no finite element simulations or pendulum tests evaluating these end terminal damage modes were conducted.”[Gabler10] Table 76. Summary of generic end terminal repair guidance. [Gabler10] Due to the importance of the end-terminal to guardrail performance, it was decided that several of these damage modes, in addition to a few others, should be quantified through

341 physical testing, including: embedment depth of foundation tubes (i.e., stub height); missing or damaged groundline strut; slack anchor cable; and rotted/weakened posts. These damage modes could affect the end-terminal’s ability to properly anchor the guardrail during impacts, which may increase the potential for excessive rail deflection, improper release of rail-post connection, and pocketing. Further, whereas local damage to a guardrail results in a relatively isolated damage-affected section of the guardrail (i.e., limited exposure to future impacts in the damage region), a damaged anchor will affect the performance of the entire guardrail system (i.e., increased risk of exposure). Research Approach The force-deflection response of a standard two-post anchor system with various types and levels of damage was determined through physical testing by applying a displacement-time history to the w-beam rail at the downstream end of the system and measuring the force- deflection response, as illustrated in Figure 287. Figure 287. Example – finite element model for computing force-deflection response of the standard two-post guardrail anchor system.[Plaxico03] Ideally, the loading on the end of the rail should be applied dynamically with a loading rate similar to real-world vehicular collisions. However, there are many difficulties involved in designing a pendulum test experiment to “pull” on the end of the system. In general, pendulums (as well as other impact devices) are used to strike and “push” an object rather than “pull” on it. Although it may be possible to construct a fixture (e.g., using pulleys or levers) to convert the compressive load of the pendulum to a tensile load on the rail, it would be difficult to separate out the dynamic effects of the mass, stiffness, energy absorption, etc. of the fixture from the test results. Also, the tension in the rail, resulting from lateral deflection of the guardrail during vehicular collisions, develops at a relatively low rate. For example, the maximum deflection of a guardrail in a TL3 event typically occurs over a period of 0.2 to 0.3 seconds. If the resulting maximum rail deflection at the anchor is, say for example, 3 inches and assuming that the rail displaces at a constant rate, then the resulting deflection rate would be in the range of 10 in/s (6.8 mph) to 15 in/s (10.2 mph). To use a pendulum (or bogie) to achieve sufficient energy to displace the anchor system at such a low rate would require a very large mass and/or a fixture

342 with significant mechanical advantage. Based on these difficulties, it was decided to use a more practical approach by performing a quasi-static test using a cable-and-winch system. A total of 10 tests were performed, in which the system was subjected to displacement- time history applied to the downstream end of the w-beam at a nominal rate of 0.5 – 1.4 in/s. The test matrix for the study is shown in Table 77 and includes evaluation of the undamaged system as well as three damage modes: (1) Missing groundline strut, (2) reduced embedment depth, and (3) slack in anchor cable. Test Set-up The tests were performed by the staff of the Federal Outdoor Impact Laboratory (FOIL) at the Federal Highway Administration’s Turner Fairbank Highway Research Center in McLean, Virginia. Figure 288 shows the typical test setup used for the study. The components of the test articles were donated by three guardrail distributors: Trinity Industries, Gregory Industries, Inc. and Road Systems, Inc. Since the purpose of these tests was to measure the force-displacement response of the anchor, only those components directly related to the anchor system were included in the test article. After each test, the components were inspected for damage. Those that showed no signs of damaged were reused in subsequent tests. Figure 289 shows a photo of the two foundation tubes with soil plates taken during the installation process. Because of the limited size of the primary soil pit, it was necessary to also utilize the adjacent soil pit to accommodate installation of the test article. This allowed sufficient distance between the posts and the edges of the soil pit for mitigating the influence of the soil pit walls on the response of the system. The distance from the front of the foundation tube at Post 1 to the front wall of Soil Pit 1 was approximately 88 inches; and the distance from the front of the foundation tube at Post 2 to the front wall of Soil Pit 2 was approximately 43 inches. The soil for all tests was classified as 21A, which conformed to Grading B of AASHTO M147-95, and was compacted in 6-inch lifts using a pneumatic tamper. The density, moisture content and degree of compaction of the soil was measured in front of and behind the post after each compaction process using a Troxler-Model 3440 Surface Moisture-Density Gauge. There were a total of twelve readings which were averaged to determine the effective soil conditions. The posts were spaced at 6.25 feet (typical). The w-beam rail was bolted to Post 2 using a 5/8-inch diameter bolt (ASTM F568) that was 10 inches long. A bracket, similar to the design used for most end-terminals, was mounted onto the end-post (Post 1), as shown in Figure 290, to allow the w-beam to slide freely in the longitudinal direction relative to the rail element. The w- beam was not bolted at the end-post. Construction drawings for the baseline generic end- terminal are provided in Appendix P.

343 Table 77. Test matrix for the anchor system damage study. Case No. FOIL Test No. Description Notes: 4B1_004 14001E Baseline System Test Undamaged generic end-terminal with two-foundation tubes, anchor cable, and groundline strut. 4B2_001 14001D No groundline strut Baseline system with groundline strut removed. 4B1_002 14001M 2 inches reduced embedment depth Baseline system with foundation tubes installed with 2 inches less embedment. 4B1_003 14001F 4 inches reduced embedment depth Baseline system with foundation tubes installed with 4 inches less embedment. 4B1_004 14001L 6 inches reduced embedment depth Baseline system with foundation tubes installed with 6 inches less embedment. 4B1_005 14001G 8 inches reduced embedment depth Baseline system with foundation tubes installed with 8 inches less embedment. 4B3_001 14001H 1 inch slack in anchor cable 4B3_002 14001I 2 inches slack in anchor cable 4B3_003 14001K 3 inches slack in anchor cable 4B3_004 14001J 4 inches slack in anchor cable Slack was measured as 1/2 the maximum vertical range of cable deflection when pushed down and pulled up using a force of 7-lbs. Measurement taken at the center of the cable length.

344 Figure 288. Typical test set-up for measuring force-deflection response of the guardrail end-terminal anchor. Figure 289. Photograph of anchor tubes with soil plates during installation (from preliminary Test 14011B). Soil Pit 1 Soil Pit 2

345 Figure 290. Photo of the simulated terminal-head bracket mounted onto the end-post. The load was applied to the end of the test article using a winch and a cable-pulley system with a 3:1 mechanical advantage, as shown in Figure 291. The winch system used for the quasi-static test was an existing component of the pendulum test device – used primarily to hoist the pendulum into position for dynamic tests. For the tests, the winch cable was run through a pulley aligned with the top corrugation of the w-beam on the test article, around another stationary pulley that was attached to a fixed bracket near the winch system, and then back to the test article and attached to the lower corrugation of the w-beam rail. For the attachment to the end of the test article, a 1-inch diameter steel rod was attached to the pulley, and another was attached to the end of the force transducer. These two rods were then fastened to the end of the w-beam rail using a standard cable-anchor-bracket (i.e., RWE02), as shown in Figure 291. A bracket-and-idler-pulley system was also mounted just downstream of the test article, as shown in Figure 292, to maintain vertical position of loading cables during the tests (e.g., simulate the continuation of the w-beam mounted to the downstream posts). Figure 291. Cable and pulley system used to apply tensile loading on end-terminal anchor. Soil Pit 1 Soil Pit 2 Simulated Terminal-Head 13,000-lb Winch25 kip Force Transducer Pulley System 3:1 Mech. Advantage Loading Brackets

346 Figure 292. Idler-pulley mount used to maintain vertical position of loading cable. Two types of posts were used for the tests. The primary differences in the two post types were that one had a round post-bolt hole with a diameter of 0.625 inches, and the other had a slotted hole 0.75 inches tall and 2.5 inches long, as shown in Figure 293. Figure 293. Comparison of the two post types used in the end-terminal tests. Idler-Pulley used to limit vertical displacement of rail Dimensions of Trinity and Gregory BCT Posts Gregory Trinity Cross-section 7.5 X 5.5 in 7.5 X 5.5 in Distance from post-bolt hole to strut-bolt hole 20.25 in 20 in Post-bolt hole size 0.625" dia. 2.5 X 0.75 in Overall length 45 in 42.75 in Post-bolt hole Strut-bolt hole Gregory Post Trinity Post

347 Equipment and Instrumentation Force Transducer The load on the cable was measured using an Interface Model 1220 standard load cell, rated at 25-kip. With the 3:1 mechanical advantage of the cable-pulley system, the resulting load on the test article was thus three times greater than the load measured by the load cell. Displacement Transducers SpaceAge Control, Inc. Series 162 Miniature Position Transducers (i.e., string-pots) were used to measure displacements at two key locations on the test article during the test, as shown in Figure 294. One string-pot was used to measure displacement at the downstream end of the w- beam rail at the load point, and another was used to measure the groundline displacement of the foundation tube at Post 1 (i.e., end post). Figure 294. Displacement transducers mounted to (a) w-beam rail at load point and (b) top of foundation tube at Post 1. Photography The tests were recorded using seven digital video cameras. Figure 295 provides the specifications and the general placement of the cameras for the test. The pre-test setup and the post-test results were also documented with “still” photographs. (a) (b)

348 Figure 295. Video camera specifications and placement.

349 Test Procedure A quasi-static displacement-time history was applied to the end of the w-beam rail of the end-terminal at a rate of approximately 0.5 – 1.4 in/sec. The test continued until there was a sudden decrease in resistance (e.g., failure of an end-terminal component) or when the displacement of the rail reached 10 inches, whichever came first. Figure 296 shows the displacement- and displacement rate-time histories measured at the load point on the end of the rail and at the groundline of Post 1 for the undamaged end-terminal case. The blue curve in Figure 297 shows the resulting force-deflection response of the end-terminal measured at the end-of the w-beam rail. The test data was resampled to obtain the “apparent” force-deflection response of the system denoted by the green curve in Figure 297. To account for the effects of loading rate on the anchor response in the finite element model, a dynamic magnification factor of 1.45 was used to scale the quasi-static force-deflection curve denoted by the red curve in Figure 297. Only two displacement transducers were available, so the displacement-time history at Post 2 was not measured in the tests. Sequential views of the undamaged end-terminal case from two view-points are shown in Figure 298. Figure 296. Displacement- and displacement rate-time histories measured at the load point on the end of the rail and at the groundline of Post 1 for baseline case. 5 The 1.4x scale factor used here was based on physical tests performed in this project (not documented here) involving static and dynamic tests on a W6x16 post in soil with density of 144 pcf. 0 2 4 6 8 10 12 14 0 1 2 3 4 5 0 1 2 3 4 5 6 7 8 9 10 11 12 D is p la ce m e n t (i n ) R at e ( in /s ) Time (sec) Displacement Rate Rail Displacement Groundline Displacement

350 Figure 297. Force-displacement response of the anchor system measured at the load point on the end of the rail for undamaged end-terminal case. Test Results A summary of the test results for Test Series 14001 are shown in Tables 78 and 79. Table 78 includes the damage mode case, installation conditions, the resulting component failures that occurred during the test, the displacement of the rail at the load point at the time of maximum load, the apparent stiffness of the anchor system during the first two inches of rail displacement, and the apparent strength of the anchor system. Table 79 includes test setup information regarding post spacing, rail height, loading point locations, embedment depth of the foundation tubes, and post-test information regarding groundline deflection of the end-terminal posts. The individual test-summary sheets are included in Appendix Q. Other than soil displacement, there were three common types of component failures that occurred during the tests: (1) Post 2 split, (2) the end-foundation tube extracted from the ground, and (3) the groundline strut buckled near the joint at Post 2. Also included in Table 78 are the results for Test 13011B, which were used to characterize the baseline anchor system model for the damaged-system-performance evaluations performed throughout the project. The primary differences in the test setup for Series 14001, compared to Test 13011B, included adding the surrogate terminal head, removal of the post-bolt at Post 1, and utilization of two soil pit areas for the installation of the post and foundation tubes (e.g., to mitigate effects of soil pit boundaries). Refer to Chapter 9 for more details regarding test setup and results of Test 13011B. The following sections discuss the effects of specific end-terminal damages on the force- deflection response of the anchor system, including the effects of missing groundline strut, reduced embedment of the foundation tubes, and slack in the anchor cable. For relative comparison, the results from the undamaged system (i.e., Test 14001E) and the baseline system (i.e., Test 13011B) are also included in the force-deflection plots for each case. 0 5 10 15 20 25 30 35 40 0 1 2 3 4 5 6 7 8 9 10 11 12 Fo rc e ( ki p s) Displacement at Rail Height (in) Test 14011E Approximate Static Estimated Dynamic (Static x1.4)

351 Figure 298. Sequential views of quasi-static test conducted for undamaged end-terminal case (Test 14011E). (0.0 in) (0.0 kips) (6.0 in) (22 kips) (2.0 in) (10 kips) (4.0 in) (19 kips) Post 2 Post 1

352 Table 78. Quasi-static test results for end-terminal damage modes (Table 1 of 2). Dry Density Moisture Compaction Dry Density Moisture Compaction 0"-2" 2"-4" 4"-6" @2" @4" @6" Peak (pcf) (%) (%) (pcf) (%) (%) (k/in) (k/in) (k/in) (k/in) (k/in) (k/in) (k) 4B1_000 13011B 12/16/2013 Baseline Test (used in Task 4A-2) 142.2 6.3 94.7 - - - 6.0 2.4 2.1 12.0 16.8 21.0 22.2 Post 1 Extraction 4B1_004 14001E 4/18/2014 Undamaged System Test 142.1 6.4 94.6 144 4.6 93.2 4.9 4.5 1.7 9.7 18.9 22.0 24.0 Soil Only 4B2_001 14001D 4/16/2014 No groundline strut 143.4 5.4 95.5 143.5 4.4 93.2 4.5 1.9 1.8 8.7 12.8 12.0 14.8 Post 2 Split 4B1_002 14001M 6/23/2014 2" reduced embedment 138.8 5.3 92.5 140.3 4.1 90.8 4.8 1.9 0.8 9.7 13.4 15.0 18.8 Post 1 Extraction 4B1_003 14001F 5/6/2014 4" reduced embedment 145.2 5.7 96.6 143.4 4.0 92.8 5.0 1.5 0.6 9.9 12.8 14.0 15.7 Post 1 Extraction 4B1_004 14001L 6/20/2014 6" reduced embedment 141.3 5.1 94.1 142.8 4.5 92.4 5.2 0.9 1.3 10.4 12.3 15.0 22.6 Post 2 Split 4B1_005 14001G 5/12/2014 8" reduced embedment 144.2 5.3 95.7 143.7 5.1 93.0 5.8 4.2 2.8 11.7 20.1 19.0 23.2 Strut Bent 4B3_001 14001H 6/4/2014 1" slack in anchor cable 143.7 5.3 95.7 143.8 3.7 93.1 5.4 4.1 2.1 10.9 19.1 23.3 29.9 Post 2 Split Strut Bent 4B3_002 14001I 6/6/2014 2" slack in anchor cable 140.4 5.7 93.5 143.4 4.3 92.8 5.4 2.8 1.3 10.8 16.4 19.0 19.3 Post 2 Split Strut Bent 4B3_003 14001K 6/18/2014 3" slack in anchor cable 140.2 5.8 93.3 144.6 4.6 93.5 2.2 3.6 2.1 4.4 11.6 15.8 22.5 Post 2 Cracked 4B3_004 14001J 6/16/2014 4" slack in anchor cable 141.8 4.6 94.4 145.1 5.3 93.9 1.3 2.3 3.5 2.6 7.3 14.2 25.2 Post 2 Split Strut Bent Stiffness Force RESULTS Test No.Case No. Damage ModeTest Date Post 1 Post 2 Soil Properties Failure

353 Table 79. Quasi-static test results for end-terminal damage modes (Table 2 of 2). Distance Distance Distance Distance Distance Height to Height to Distance Post 2 to Soil Pit Front Post 2 to Soil Pit Rear Post 1 to Soil Pit Front Post 1 to Soil Pit Rear Center to Center Posts Load Point Top Load Point Bottom Center to Center Posts Post #1 Post #2 (in) (in) (in) (in) (in) (in) (in) (in) (in) (in) (in) (in) 13011B 12/16/2013 Baseline Test (used in Task 4A-2) 14001E 4/18/2014 Undamaged System Test 88.0 16.0 43.0 30.0 75.5 28.5 25.5 17.5 57.0 74.0 5.8 4.0 14001D 4/16/2014 No groundline strut 86.0 18.0 45.0 28.0 74.0 - - - - 70.0 4.3 < 0.125 14001L 5/20/2014 2" reduced embedment 88.5 17.8 42.0 30.5 75.3 29.5 26.5 18.5 55.0 75.0 8.0 7.0 14001F 5/6/2014 4" reduced embedment 85.8 18.3 41.3 31.8 74.5 31.5 28.5 17.3 53.0 73.3 4.8 4.0 14001L 6/20/2014 6" reduced embedment 87.0 19.3 42.0 30.5 76.5 34.5 31.5 23.5 51.0 73.3 9.0 6.5 14001G 5/12/2014 8" reduced embedment 85.5 18.5 44.5 28.5 75.0 31.0 28.0 20.0 49.0 71.5 4.5 3.0 14001H 6/4/2014 1" slack in anchor cable 87.5 18.3 41.5 25.0 75.3 26.0 23.0 16.0 57.0 69.0 9.5 3.0 14001I 6/6/2014 2" slack in anchor cable 86.8 19.5 40.8 31.8 75.5 26.5 22.8 15.0 57.0 69.5 9.8 5.0 14001K 6/18/2014 3" slack in anchor cable 87.0 19.3 42.0 30.5 75.0 29.0 26.0 18.0 57.0 72.8 8.0 6.5 14001J 6/16/2014 4" slack in anchor cable 86.8 19.5 40.8 31.8 76.3 27.3 24.3 17.0 57.0 68.8 9.5 3.0 Groundline Deflection Height to Top of Rail PRE-TEST MEASUREMENTS POST-TEST RESULTS Test No. Test Date Damage Mode Post Embedment Depth

354 Missing or Non-Functioning Groundline Strut Pre-test and post-test photos of Test 14011D are shown in Figures 299 and 300, respectively. The data from the test were resampled to obtain the “approximate static” force- deflection response, which is compared to that of the undamaged system in Figure 301. Refer to the test-summary sheets in Appendix Q for the “unprocessed” force-deflection results. Figure 302 shows the displacement-time history measured at the load-point on the rail and at the groundline of Post 1. The “apparent” initial stiffness6 of the anchor was 4.5 kips/in for the first 2 inches of displacement measured at the load point, compared to 4.9 kips/in for the undamaged system. Between 2-4 inches displacement the stiffness dropped to 1.9 kips/in, compared to 4.5 kips/in for the undamaged system. Post 2 split at approximately 5 inches displacement at which point the loading on the system was 14.8 kips resulting in a sudden loss of the anchor resistance. The test was terminated immediately after the loss of Post 2 with a maximum displacement of 6.4 inches measured at the load point; thus, the response of the system beyond this displacement was not determined. The apparent strength of the anchor was 14.8 kips and occurred at 5 inches deflection. Figure 299. Pre-test photo of Test 14001D (front view). 6 Apparent initial stiffness was computed as the force at 2 inches displacement divided by 2 inches; the apparent stiffness from 2-4 inches was calculated as the difference in force from 2-4 inches divided by 2 inches; etc.

355 Figure 300. Post-test photo of Test 14001D (back view). Figure 301. Force vs. deflection response for the standard end-terminal with missing groundline strut. Post 1Post 2 0 5 10 15 20 25 30 0 2 4 6 8 10 12 Fo rc e ( ki p s) Longitudinal Deflection at Rail Height (in) No Groundline Strut (Approximate Static) Baseline (Test 13001B) Undamaged (Test 14001E) Missing Strut (Test 14001D)

356 Figure 302. Displacement and displacement rate vs. time for Test 14001D. The results for the test involving the missing groundline strut showed that the initial stiffness of the end-terminal was similar to that of the undamaged system for approximately the first 1.5 inches of longitudinal rail movement. This was because the length of the groundline strut was 67 inches, compared to the distance of 68 inches measured between the two foundation tubes for the undamaged test case (refer to Test 14001E in Table 79). This resulted in a total gap of 1.0 inch between the strut and foundation tubes, as shown in Figure 303. Thus, the foundation tube at Post 1 during the test on the undamaged system experienced 1 inch of movement before the groundline strut engaged the foundation tube at Post 2. Figure 303. Typical position of groundline strut in a standard two-post-strut anchor system relative to (a) Post 2 and (b) Post 1. Reduced Embedment Depth A total of four tests were performed to quantify the effects of reduced embedment depth on the force-deflection response of the standard anchor system. These tests included installations with 2, 4, 6 and 8 inches of reduced embedment depth. At the standard embedment depth for the system, the top of the foundation tube protrudes 3 inches above the ground. Thus, amount of reduced embedment was determined by measuring the distance from the top of the soil to a point three inches from the top of the foundation tube at Post 1. Figure 304 shows an example of the procedure for the case of 6 inches reduced embedment. 0 2 4 6 8 10 12 14 0 1 2 3 4 5 0 1 2 3 4 5 6 7 8 9 10 11 12 D is p la ce m e n t (i n ) R at e ( in /s ) Time (sec) Displacement Rate Rail Displacement Groundline Displacement (a) (b)

357 Figure 304. Procedure for measuring reduced embedment depth. Pre-test photos for the four reduced-embedment cases are shown in Figures 305 through 308, and post-test photos are shown in Figures 309 through 312. The “approximate static” force- displacement results from the tests are shown in Figure 313, including the undamaged cases. The results from this series of tests were somewhat inconclusive. The initial stiffness (i.e., between 0- 2 inches displacement) for all reduced embedment cases was essentially identical to the undamaged case (i.e., 4.9 kips/in). The stiffness at deflections of 2 to 4 inches, however, was significantly lower for reduced embedment cases of 2, 4 and 6 inches, and reduced slightly for each successive reduction of embedment (i.e., 1.9, 1.5 and 0.9 kips/in, respectively, compared to 4.5 kips/in for the undamaged case). For the 8-inch reduced embedment depth, on the other hand, the stiffness at this second level of displacement was essentially the same as the undamaged case (e.g., 4.2 vs. 4.5 kips/in). For the case involving 2 inches of reduced embedment of the foundation tubes, the “apparent” initial stiffness of the anchor was 5.0 kips/in for the first 2 inches of displacement measured at the load point, compared to 4.9 kips/in for the undamaged system. Between 2-4 inches displacement the stiffness dropped to 1.5 kips/in, compared to 4.5 kips/in for the undamaged system. Between 4-6 inches displacement the stiffness dropped further to 0.6 kips/in, compared to 1.7 kips/in for the undamaged case. The maximum load measured during the test was 15.7 kips at 10.5 inches displacement. Post 1 foundation tube began to extract from the ground at a rail deflection of approximately 7.3 inches. At approximately 13 inches displacement the foundation tube at Post 1 was fully extracted. Refer to the test-summary sheets in Appendix Q for more details. For the case involving 4 inches of reduced embedment of the foundation tubes, the “apparent” initial stiffness of the anchor was 4.8 kips/in for the first 2 inches of displacement measured at the load point. Between 2-4 inches displacement the stiffness dropped to 1.9 kips/in. Between 4-6 inches displacement the stiffness dropped further to 0.8 kips/in. The maximum load

358 measured during the test was 18.4 kips at 16 inches displacement. Post 1 foundation tube began to extract from the ground at a rail deflection of approximately 3 inches (based on assessment of real-time video and force-displacement data). Refer to the test-summary sheets in Appendix Q for more details. For the case involving 6 inches of reduced embedment of the foundation tubes, the “apparent” initial stiffness of the anchor was 5.2 kips/in for the first 2 inches of displacement measured at the load point. Between 2-4 inches displacement the stiffness dropped to 0.9 kips/in. Between 4-6 inches displacement the stiffness increased to 1.3 kips/in. The maximum load measured during the test was 22.6 kips at 12 inches displacement. Post 2 split at approximately 5.1 inches of rail deflection. The rail then dropped at Post 2. The foundation tube at Post 1 showed no signs of pulling out of the ground for this case. Refer to the test-summary sheets in Appendix Q for more details. For the case involving 8 inches of reduced embedment of the foundation tubes, the “apparent” initial stiffness of the anchor was 5.8 kips/in for the first 2 inches of displacement measured at the load point. Between 2-4 inches displacement the stiffness was 4.2 kips/in. Between 4-6 inches displacement the stiffness was 2.8 kips/in. The maximum load measured during the test was 23.2 kips at 5.0 inches displacement. The foundation tube at Post 1 began to extract from the ground at approximately 5 inches rail deflection. At approximately 8 inches deflection the groundline strut buckled near Post 2, resulting in a sudden loss of anchor resistance. Refer to the test-summary sheets in Appendix Q for more details. Figure 305. Pre-test photo of Test 14001M for 2-inch reduced embedment case.

359 Figure 306. Pre-test photo of Test 14001F for 4-inch reduced embedment case. Figure 307. Pre-test photo of Test 14001L for 6-inch reduced embedment case.

360 Figure 308. Pre-test photo of Test 14001G for 8-inch reduced embedment case. Figure 309. Post-test photo of Test 14001M for 2-inch reduced embedment case.

361 Figure 310. Post-test photo of Test 14001F for 4-inch reduced embedment case. Figure 311. Post-test photo of Test 14001L for 6-inch reduced embedment case.

362 Figure 312. Post-test photo of Test 14001G for 8-inch reduced embedment case. Figure 313. Force vs. deflection response for the standard end-terminal with 0 to 8 inches reduced embedment depth of foundation tubes. 0 5 10 15 20 25 30 0 2 4 6 8 10 12 Fo rc e ( ki p s) Longitudinal Deflection at Rail Height (in) Embedment Depth (Approximate Static) Baseline (Test 13001B) Undamaged (Test 14001E) 2" Less Embed (Test 14001M) 4" Less Embed (Test 14001F) 6" Less Embed (Test 14001L) 8" Less Embed (Test 14001G)

363 Slack Anchor-Cable A total of four tests were performed to quantify the effects of slack in the anchor-cable on the force-deflection response of a standard anchor system. These tests included installations with 1, 2, 3 and 4 inches of slack in the anchor cable. The procedure for measuring the amount of slack in the cable involved applying a downward vertical force of 7 lbs. at the center of the cable length and then measuring the vertical distance from the top of the cable to the ground. The cable was then pulled upward with a vertical force of 7 lbs. and again the distance from the top of the cable to the ground was measured. The amount of slack in the cable was then calculated as the difference in these two values divided by 2. In other words, slack was defined as one-half the maximum vertical range of cable’s deflection when pushed down then pulled up using a force of 7-lbs, as shown schematically in Figure 314. This definition of slack is consistent with that specified by most manufacturers of proprietary end-terminals. The 7-lb force is representative of the amount of force required when applying reasonable pressure with one’s index finger to push down or pull up on the cable. When performing this procedure, care was taken to ensure that all the slack was taken out of the cable by visually confirming that the bearing plate on the front of Post 1 was seated firmly against the post under the applied pressure. Figure 314. Illustration showing procedure for measuring slack in cable-anchor. Pre-test photos for the four slack-cable cases are shown in Figures 315 through 318, and post-test photos of are shown in Figures 319 through 322. The “approximate static” force- displacement results from the tests are shown in Figure 323, including the undamaged cases. The results show that the initial force-deflection response of the end-terminal was essentially unaffected until the slack in the cable exceeded 2 inches. At 3 inches of slack, there was a significant drop in the force-deflection response, particularly during the first 1-inch of displacement. This was partly due to the fact that Post 2 for that test included a 2.5-inch long slotted post-bolt hole (rather than the 0.625” diameter post-bolt hole for all other tests). The slotted hole effectively delayed the point at which the post-bolt engaged the edge of the hole and began to load the post. The end result was that Post 2 was able to survive significant rail displacement without failure, but also with the effect of slightly reducing the initial stiffness of the anchor system.

364 Figure 315. Pre-test photo of Test 14001H for 1” slack-cable case. Figure 316. Pre-test photo of Test 14001I for 2” slack-cable case.

365 Figure 317. Pre-test photo of Test 14001K for 3” slack-cable case. Figure 318. Pre-test photo of Test 4001J for 4” slack-cable case.

366 Figure 319. Post-test photos of Test 14001H for 1” slack-cable case. Figure 320. Post-test photos of Test 14001I for 2” slack-cable case. Figure 321. Post-test photos of Test 14001K for 3” slack-cable case. Post 1Post 2 Post 1Post 2 Post 1Post 2

367 Figure 322. Post-test photos of Test 14001J for 4” slack-cable case. Figure 323. Force vs. deflection response for the standard end-terminal with 0 to 4 inches of slack in the anchor-cable. For the case involving 3 inches of slack in the cable, the “apparent” initial stiffness of the anchor was 2.2 kips/in for the first 2 inches of displacement measured at the load point, compared to 4.9 kips/in for the undamaged system. Between 2-4 inches displacement the stiffness increased to 3.6 kips/in, compared to 4.5 kips/in for the undamaged system. Between 4- 6 inches displacement the stiffness dropped back to 2.1 kips/in, compared to 1.7 kips/in for the undamaged case. The maximum load measured during the test was 22.5 kips at 12 inches displacement. No significant failures occurred during the test. Refer to the test-summary sheets in Appendix Q for more details. For the case involving 4 inches of slack in the cable, the “apparent” initial stiffness of the anchor was 1.3 kips/in for the first 2 inches of displacement, compared to 4.9 kips/in for the Post 1Post 2 0 5 10 15 20 25 30 35 0 2 4 6 8 10 12 Fo rc e ( ki p s) Longitudinal Deflection at Rail Height (in) Slack Anchor Cable (Approximate Static) Baseline (Test 13001B) Undamaged (Test 14001E) 1" Slack Cable (Test 14001H) 2" Slack Cable (Test 14001I) 3" Slack Cable (Test 14001K) 4" Slack Cable (Test 14001J)

368 undamaged system. Between 2-4 inches displacement the stiffness increased to 2.3 kips/in, compared to 4.5 kips/in for the undamaged system. Between 4-6 inches displacement the stiffness increased to 3.5 kips/in, compared to 1.7 kips/in for the undamaged case. The maximum load measured during the test was 25.2 kips at 10 inches displacement. Post 2 split at 3.3 inches of rail displacement at 5.8 kips, and the groundline strut buckled at 10.8 inches of rail deflection at 20.7 kips. Refer to the test-summary sheets in Appendix Q for more details. Summary and Discussion Although the end-terminal of a guardrail serves many purposes, one of its primary functions is to “anchor” the ends of the rail so that the resulting tension in the rail can help to limit lateral deflection of the system during impacts. The objectives of this study were to (1) quantify the effects of various types and levels of end-terminal damage on the force-deflection response of standard generic guardrail end-terminals using physical testing, and (2) to then correlate those effects to guardrail performance. In Chapter 9 the response of the G4(2W) guardrail system was evaluated for various levels of reduced anchor strength. In the current study quasi-static tests were performed to quantify the force-deflection response of a standard two-post-and-strut anchor system with three modes of damage: (1) missing or otherwise non- functional groundline strut, (2) reduced embedment depths of 2, 4, 6, and 8 inches, and (3) slack anchor cable with 1, 2, 3, and 4 inches of slack. The tests were performed by pulling on the downstream end of the rail element with a cable and winch system at a displacement rate of 0.5 to 1.4 in/s. The force-deflection response was measured via a force transducer and string-pot which were attached to the rail at the load point. The following is a brief summary and discussion of the results. Discussion of Failure Modes There were basically three types of anchor component failures that occurred during the test series: (1) Post 2 splitting, (2) the foundation tube at Post 1 pulling out of the ground, and (3) buckling of the groundline strut. Based on the test results, however, there was no conclusive evidence that the failure of Post 2 in any way degraded anchor performance; and is therefore not considered a failure mode for the system. In fact, in all cases for which the test was allowed to continue after Post 2 failed, the effective stiffness of the system remained more or less continuous, as illustrated in Figure 324 (also see results for Tests 14001L, 14001H, and 14001J in Appendix Q). In field installations the failure of Post 2 would, however, be an indication that other aspects of the system may be compromised, such as significant movement of the anchor foundation tubes or slack in the anchor cable. For cases in which the foundation tube at Post 1 began pulling out of the ground, there was noticeable reduction in stiffness of the system during the extraction process (e.g., refer to Tests 14001M, 14001F, and 14001G). This failure mode was primarily associated with reduced embedment of the foundation tubes. When the foundation tube was fully extracted there was an abrupt loss of the anchor. There was also an abrupt loss of anchor resistance for cases in which the groundline strut buckled. Although there was one case of this failure mode occurring for the reduced embedment cases, it was principally associated with slack in the anchor cable, and generally occurred when groundline deflections at Post 1 reached approximately 7-8 inches.

369 Figure 324. Test 14001H showing continuing effectiveness of anchor after Post 2 fails. Missing Groundline Strut The results of the tests showed that the missing groundline strut resulted in a reduction in anchor capacity of approximately 50%; however, the initial stiffness of the anchor was very similar to the undamaged case. Also, the test was terminated prematurely (i.e., soon after Post 2 split), thus the full range of effective anchor response was not measured for this case. Regarding the effects on guardrail performance, it is expected that the damage mode involving a missing (or otherwise non-functional) groundline strut would be similar to that of the FE analysis for the 47% percent anchor strength case in Chapter 9. Recall from Chapter 9 that when end-terminal damage results in more than 50% reduction in anchor capacity, the recommended repair priority was “high”. This was particularly true for the G4(1S), due to the low torsional strength of the W6x9 steel posts; and for wood post guardrail systems when the posts had a degradation level of DL1 or higher. For example, the analysis of the G4(1S) in Chapter 11, which involved an end-terminal with no groundline strut and 14 inches of pre-crash-induced rail deflections, the upstream anchor deflected 5.5 inches, which lead to the vehicle overriding guardrail. Test C080C3-027-2 of this same system, also resulted in excessive anchor movement and vehicle override. Reduced Embedment Depth for the Foundation Tubes The results from this series of tests were somewhat inconclusive. The initial stiffness (i.e., corresponding to displacements from 0-2 inches) on the anchor for all of the reduced embedment cases was essentially identical to that of the undamaged case (i.e., 4.9 kips/in). The stiffness at deflections of 2 to 4 inches, however, was significantly lower for the reduced embedment cases and tended to progressively decrease as a function of embedment depth, with the exception of the 8-inch reduced embedment case. The effective stiffness values were 1.9, 1.5, 0.9, and 4.2 kips/in for the 2”, 4”, 6” and 8” reduced embedment cases, respectively. The stiffness at this second level of displacement for the 8-inch reduced embedment case was, unexpectedly, of similar magnitude as for the undamaged case (i.e., 4.5 kips/in). 0 5 10 15 20 25 30 35 40 0 1 2 3 4 5 6 7 8 9 10 11 12 Fo rc e ( ki p s) Displacement at Rail Height (in) Test 14001H Approximate Static Estimated Dynamic (Static x1.4) Post 2 Failed

370 In NCHRP Report 656, Gabler et al. stated that when the stub height above ground level exceeds 4 inches (i.e., reduced embedment of 1 inch) then the breakaway mechanism of the end- terminal may not activate properly during end-on hits.[Gabler10] Gabler further stated that stub heights exceeding 4 inches could lead to small vehicles snagging on the top of the foundation tube. Based on those observations, Gabler recommended that stub heights of more than 4 inches above ground be assigned a repair priority of “medium”. In those assessments, however, only the performance of the end-terminal in regard to direct impact on the end of the system was considered. Whereas, in the current study, the focus was on the effects of reduced embedment (i.e., or stub height) with regard to the force-deflection response of the anchor, which would thus affect the entire length of need of the guardrail system. From the evaluations of the effects of varying anchor strengths on the performance of the G4(2W) guardrail in Chapter 9, it was determined that for undamaged wood posts (i.e., DL0) the performance of the system was not significantly affected by anchor strength. In those cases, the torsional rigidity of the line posts effectively allowed the posts to carry a large proportion of the tensile load in the guardrail. However, when the deterioration level for the posts were DL1 or higher, the performance of the system was compromised when anchor strength was reduced by 30 percent (relative to the baseline case). From the tests performed in this study, the response of the anchor with 2 inches reduced embedment indicated negligible reduction in strength for the first 2 inches of rail deflection, and only 16 percent reduction in strength at 6 inches deflection. Beyond 6 inches deflection the reduction in strength gradually approached 31 percent, but it is expected that the system would have provided sufficient anchorage to successfully contain and redirect most vehicle impact situations prior to reaching such magnitudes of deflection. The response of the anchor with 4 inches reduced embedment was similar to that at 2 inches reduced embedment, but for this case the resistance of the system at approximately 3 inches of displacement was 13 kips, and remained at this level up to approximately 10 inches displacement. The response at 6 inches reduced embedment began to experience lower resistance at approximately 2 inches deflection, but not significantly less than that of the 4 inches reduced embedment case. At 8 inches reduced embedment, the initial response was undistinguishable from that of the undamaged case, but the loss of anchor resistance occurred abruptly at only five inches rail displacement, which corresponded to 2.6 inches groundline deflection. Based on the results of these tests combined with those from Gabler’s study, it was concluded that:  When the “stub height” of the foundation tube, as measured from the top of the foundation tube to the ground, is greater than 4 inches and less than 9 inches (i.e., reduced embedment of 1 to 6 inches), repair should be considered if other system maintenance is being performed. The performance of the anchor is not compromised when stub height is less than 7 inches; however, there is an increased potential for small cars snagging on the top of the foundation tube as well as an increased potential for the breakaway mechanism of the end-terminal to not activate properly during end-on hits at stub heights exceeding 4 inches.  When stub height exceeds 9 inches (i.e., reduced embedment exceeds 6 inches) the performance of the anchor may be significantly compromised due to significant reduction in anchor strength and the high probability of complete anchor loss at relatively low anchor displacement.

371 Slack in Anchor Cable For the damage mode defined by slack in the anchor cable, the tests showed that the force-deflection response of the end-terminal was essentially unaffected for up to approximately 6 inches of rail deflection for cases where the initial slack in the cable was two inches or less. At 3 and 4 inches of slack, however, there was a significant drop in the force-deflection response. In fact, the resistance during the first 1 inch of displacement was essentially negligible for both of these cases. At two inches of displacement the effective resistance was reduced by 65 percent for the case of 3 inches slack, and reduced by 73 percent for the case of 4 inches slack, relative to the baseline case. At four inches of displacement the effective resistance was reduced 39 percent for the case of 3 inches slack; and 62 percent for the case of 4 inches slack. For these two test cases the ultimate strength of the anchor tended to approach a similar magnitude as that of the undamaged case as the displacement approached 10 inches. During a crash event, however, the reduced tension in the rail at the early stages of the crash would likely have resulted in excessive lateral deflections and possibly severe pocketing of the guardrail system by the time the anchor became effective. Based on the results from the evaluations of anchor strength on the performance of the G4(2W) guardrail in Chapter 9, when the slack in the anchor cable is greater than 2.0 inches and less than 3 inches the performance of the anchor system may be somewhat compromised; and when the slack in the cable exceeds 3 inches, the performance of the anchor may be significantly compromised. According to manufactures recommendations, the anchor cable should be tightened when there is more than 1 inch of slack. When possible this simple repair should be performed during the inspection process. Recommendations The following recommendations are based on the collective results from (1) evaluations of the G4(2W) guardrail with varying anchor strength (i.e., Chapter 9), (2) the force-deflection response of the standard two-post-and-strut anchor subjected to tensile loading and (3) the recommendations by Gabler et al. in NCHRP Report 656.[Gabler10] The analysis results from Chapter 9 indicated that the performance of the G4(2W) with healthy posts (i.e., DL0 or DL1) was not significantly affected by anchor strength for the case of an impact on the system at approximately 62 feet downstream of the anchor – this impact point was adopted from full-scale Test 471470-26 which was used as the baseline system throughout this study. In this impact scenario, there were enough posts upstream of the impact point carrying the tensile load of the rail such that the loading on the anchor was negligible. However, consideration should be given to the possibility of impact occurring at a point nearer to the anchor which would significantly increase the loading on the anchor. Also, if the w-beam rail detaches from the upstream posts during impact, or if posts upstream of the impact point fracture under the tensile loading of the w-beam, then anchor loads may increase significantly. This result was demonstrated in this study by simulating Report 350 Test 3-11 impact on the G4(2W) in which the anchor strength was reduced by 47 percent (e.g., end-terminal with single foundation- tube anchor) and the posts were modeled with deteriorated strength properties. The results indicated that as post deterioration levels increased, there was an increase in the number of posts upstream of the impact that fractured, resulting in increased loading on the anchor.

372 As a result of this study, the research team recommends that the repair threshold for the end-terminal for the G4(2W) be those in which the damage results in more than a 30% loss in anchor capacity relative to the baseline anchor strength. When the damage results in more than 50% loss of capacity then the relative priority for repair is high. These cases include:  High Priority: o Missing or otherwise non-functional groundline strut, o Reduction in embedment depth exceeding 6 inches for foundation tubes (i.e., stub height exceeding 9 inches), and o Anchor cable slack exceeding 3 inches. For end-terminal damage that results in 30% to 50% loss of capacity for the anchor, the relative priority for repair is medium, unless a fixed object is located within 50 inches behind the face of the barrier. In that case, the line posts should also be checked for deterioration damage. If the damage level for the posts is DL1 or greater, then the priority for repair is high, based on lateral deflection limitations of the guardrail. These cases include:  High Priority: o Fixed/rigid object located within 50 inches of guardrail, reduced embedment of foundation tube exceeding 4 inches (i.e., stub height exceeding 7 inches) and line posts with deterioration level of DL1 or greater.  Medium Priority: o No fixed objects located within 50 inches of guardrail, reduced embedment of foundation tube exceeding 4 inches but less than 6 inches (i.e., stub height protruding 7 to 9 inches above grade). A summary of the recommendations regarding end-terminal damage for the G4(2W) guardrail are presented in Table 80. For completeness, the recommendations from NCHRP Report 656 for damage modes involving the end post, cable bracket, lag bolts on impact head, and bearing plate are also included. The recommendations derived from the current study pertain, primarily, to the upstream anchor. The downstream end-terminal does not generally experience the same magnitude of loading as does the upstream terminal. For this reason, many guardrail systems use a standard “trailing end terminal” (i.e., a single foundation tube and anchor cable) at the downstream end of the guardrail. For undivided highways, where the impact on a guardrail may occur from either the primary or opposing directions of traffic, the anchor at either end of the guardrail should meet the recommendations of Table 80. For divided roadways, on the other hand, guardrails are only exposed to impacts from traffic in one direction (the primary direction), thus the downstream anchor will require less stiffness/strength compared to the upstream anchor.

373 Table 80. Recommendations for end-terminal damage for the G4(2W). Future Work  Test Instrumentation o A displacement transducer should be installed on the upstream end of the rail to provide a more accurate representation of the longitudinal displacement. The problem with measuring displacement at the load point is that there are many cases when there is significant lateral and vertical displacement of the rail (e.g., when Post 2 breaks and the rail suddenly drops). The displacement transducer only measures the change in length of the “string” and thus can’t discern direction of motion. Damage Mode Repair Threshold Relative Priority Damaged End Post High Med Cable Anchor Bracket Medium - High - Medium Groundline Strut High High Med Lag Bolts (Energy Absorbing Terminals Only) High - High - High - End-Terminal damage that results in more than 30% loss of anchor capacity and - Guardrail line-posts with deterioration levels of DL1 or greater. These cases include: - Reduced embedment exceeding 4 inches (i.e., stub height exceeding 7 inches) and line posts with deterioration level of DL1 or greater. Not functional (Sheared, rotted, severely cracked) Loose or misaligned Greater than 2 inches and less than 3 inches End-Terminal damage that results in more than 50% reduction in anchor capacity (relative to the baseline anchor strength). Stub height exceeds 9 inches (e.g., soil plate is visible above grade) Stub height exceeds 4 inches but less than 9 inches Other Foundation Tube Anchor Cable Bearing Plate High Missing or otherwise nonfunctional Missing Missing or failed If a hazard is located within 50 inches behind the w-beam rail,and includes the combination damage mode of: Missing More than 3 inches slack. Loose or not firmly seated in rail

374  Crash Performance Evaluations o Future work should include analyses of the G4(2W) at impact points nearer to the anchor system. o Due to the differences in the torsional rigidity of the W6x9 steel posts of the G4(1S) guardrail, an investigation similar to the one conducted in Chapter 9 should also be conducted to assess effects of anchor strength on the performance of that system, since it is expected that the G4(1S) would have a greater sensitivity to anchor strength.